UKF位姿估计的超静平台耦合模型参数辨识

Parameter Identification for Ultraquiet Platform Coupling Model Based on UKF

针对载荷无陀螺时辨识超静平台耦合动力学参数存在位姿确定问题,设计了一种基于Schur分解以及无迹卡尔曼滤波（UKF）的位姿确定及参数辨识方法。首先,建立加速度计和姿态敏感器组成的测量系统状态模型和观测模型,并给出测量系统的可观性分析。然后,给出基于UKF的载荷位姿确定方法;在UKF中引入姿态修正信息,从而提高载荷角速度估计精度,实现载荷广义位移、广义速度、广义加速度的准确估计。通过Schur分解实现超静平台动力学模型解耦及辨识模型中动力学参数显式表达。以滤波器估计载荷位姿信息为依据,采用最小二乘法辨识动力学参数。仿真结果表明UKF能够准确估计载荷角速度以及超静平台支杆刚度系数,辨识误差优于百分之一。

A method based on Schur decomposition and unscented Kalman filter（UKF） is designed to identify the parameters of the Stewart coupling model. The state model,measurement model and observability are given. The UKF is adopted to determine the payload position and attitude. The attitude measurements are used in the UKF to correct the angular velocity error so that the estimation accuracy and convergence speed of the angular velocity can be improved. The payload generalized displacement,generalized velocity and generalized acceleration should be obtained. The parameter decoupling method is designed and the least squares method is used to identify the dynamics parameters. Numerical simulations results indicate that the angular velocity is estimated accurately by the UKF and the stiffness identification errors of the Stewart six struts are less than 0. 01.